Product sense

Product Sense Skill

Apply a structured framework to PM product sense / product design questions targeting AI product roles.

When to Use

• User asks a "Design a product for X" question
• User asks "How would you improve X"
• User asks "How would you productize X capability"
• User says /product-sense followed by a question
• Any product design, product sense, or "build a product" interview question

Context

• Tuned for: AI product roles at frontier AI companies
• What matters: First-principles thinking, ambition, structured clarity, and taste
• Common pitfall: Rushing to solutions without clarifying the problem first. Always start with clarifying questions.

Framework: Product Sense (6 Sections)

Generate the answer following this EXACT structure. Each section should be substantive - not just headers.

Section 1: Clarifications (ASK FIRST, ALWAYS)

Ask 3-5 clarifying questions before proceeding. Categories:

• Scope: What company are we? What's the form factor? Platform constraints?
• Users: Who is the primary audience? B2C vs B2B vs B2B2C?
• Business: What stage is the company? Revenue model? Strategic priorities?
• Technical: What capabilities exist? What's feasible in the timeframe?
• Constraints: Budget, timeline, regulatory, geographic?

After listing questions, state reasonable assumptions for each and proceed.

Section 2: Product Strategy & Rationale (WHY BUILD THIS)

• Company Mission: How does this align with the company's stated mission? Reference the specific company's mission statement and connect your product thinking to it.
• Trends & Tailwinds: What macro trends make this timely? (AI adoption curves, regulatory shifts, user behavior changes)
• Competition: Who else is doing this? What's the gap?
• Strategic Moat: What unique advantage does this company have here?
• Product Goal: One sentence on what we're building and why NOW

Section 3: User Segmentation (WHO)

Segment users along 3 dimensions and pick a primary:

• Reach: How many potential users in each segment?
• Frequency: How often would they use this?
• Underserved: How poorly served are they today?

For each segment provide: persona name, description, why they'd use this, current alternatives.

Pick primary segment with clear rationale (usually: highest frequency + most underserved).

Section 4: User Journey & Pain Points (WHAT HURTS)

Map the user journey for the primary segment:

1. Discovery/Awareness
2. Onboarding/First Use
3. Core Usage Loop
4. Retention/Return

For each stage, identify pain points scored on:

• Severity (1-3): How painful is this?
• Frequency (1-3): How often does it occur?
• Alternatives (1-3): How well do current solutions address this?

Pick top 2-3 pain points to solve. Justify the prioritization.

Section 5: Solutions (HOW)

For each selected pain point:

Brainstorm (3-4 options per pain point, range from incremental to ambitious)

Evaluate each on:

• Impact on pain point (High/Med/Low)
• Engineering effort (High/Med/Low)
• Strategic alignment (High/Med/Low)
• Differentiation (High/Med/Low)

Recommend top solution with clear rationale. Describe:

• What the user sees/experiences (be concrete and specific)
• Key features for V1 vs V2
• Why this is better than alternatives

Section 6: Success Metrics

• North Star Metric: One metric that captures the core value delivered
• Supporting Metrics (3-4): Leading indicators that the NSM will grow
• Counter/Guardrail Metrics (2-3): What we must NOT break (safety, quality, trust)
• How to measure: What instrumentation is needed?

AI Company Flavor

When the question is about an AI company product, layer in:

• Safety considerations: How does this product avoid harm? What guardrails exist?
• Model capabilities: What model capabilities enable this? What's the technical frontier?
• Scaling dynamics: How does this get better with more users/data?
• Mission alignment: Tie back to the specific company's mission
• Taste: Match the company's culture — some labs value ambitious, creative, step-change thinking; others value careful, principled, safety-first thinking; others value scientific rigor. Research the specific company's values beforehand.

Output Format

Structure the response as a conversational walkthrough — structured but natural. Use the section headers. Aim for ~2500 words total. Flag where you'd pause for discussion or input.

Research-First Workflow

Before generating the answer:

1. Research — Use web search to find latest thinking from AI company blogs, PM thought leaders, market data, competitor intel. Do 5-10 searches.
2. Cite sources — Include [linked source](url) inline for major claims, data points, and trends.
3. Display the complete structured answer.

What Good Looks Like

• Starts with clarifying questions (CRITICAL)
• Shows strategic thinking before jumping to solutions
• User empathy is specific and grounded (not generic)
• Solutions are creative AND feasible
• Metrics are specific and tied to user value
• Mentions trade-offs and what you'd NOT build
• Ties back to company mission
• Shows taste and opinion (not just framework execution)